Exploring the role Ocado's AI plays in keeping food waste low across grocery supply chains

Ask any retail or logistic operator what the biggest complexity in their operation is, and you’ll find that supply chain challenges come up time and time again. In no industry is this more apparent than online grocery, where retailers must navigate a sea of changing factors, from fluctuating demand to freshness, accuracy, and customers’ ever-growing need for product availability and wider choice.

Ocado’s pioneering technologies are changing this reality – bringing unforeseen levels of visibility into global grocery supply chains to help retailers maximise their customer proposition and efficiency, while keeping waste low. 

Through our integrated technology platform, we’re helping grocers keep track of inventory, save costs, and make the most out of every opportunity. 

Why is supply chain complexity such a challenge for online grocers?

Traditional food supply chains are ripe with inefficiencies due to overly complicated processes and far too many third-party providers plugged into different processes with complicated integrations.

In this environment, retailers may face a number of trade-offs. Do they prioritise low food waste above decisions on merchandising or product range, for example? Or how do they weigh up early profits over long-term benefits?

At a basic level, this trade-off looks like:

1. A retailer not ordering enough stock - The customer doesn’t receive what they ordered and satisfaction levels decrease. 
2. A retailer ordering too much stock - Customers may be happier but profitability is lower and waste is higher.

When it comes to striking a balance, preferences vary. One retailer may prioritise availability and fulfilment accuracy for customers, while another may want to drive down waste. The challenge is finding the sweet spot.

How is Ocado solving supply chain challenges while keeping food waste low? 

Ocado’s technology empowers global grocery retailers to take ownership of their supply chains. Through the embedded intelligence of the Ocado Smart Platform (OSP) – our configurable end-to-end ecommerce, fulfilment and logistics solution – retailers are leveraging data to drive stronger economics.

Our smart fulfilment systems can accurately detect the risk of purge from any site on any given day, at any given time. This is only possible because of the fully integrated software applications and AI powering OSP.‍

Let’s take bananas as an example. OSP’s intelligence can account for: 

• Stock - How many bananas have already been assigned to customers? How many are still in stock, and how many are due to come in?
• Quality - What is the shelf-life of these bananas? 
• Demand - What is the anticipated demand for bananas?

With this information, we can accurately forecast sales and purge using machine learning modelling to hit the sweet spot between product availability and customer demand.

Lowering waste through smart inventory management 

In many traditional online fulfilment operations, customers search for products before picking a slot. On OSP, the slot is the first step. Customers choose their delivery time, day and location before they’ve even had a chance to browse any product.

This change is crucial to achieving order accuracy. Our advanced inventory management can make real-time predictions on the types of products to show customers on the webshop because it knows what products are currently available, what orders are due to be coming in from suppliers, how long products will stay fresh, and what can be ordered in time to make the delivery. This enhanced visibility helps to keep waste low. 

Enhanced planning and forecasting through one connected system 

Today’s retailers operate multiple warehouses and webshops simultaneously with vast product ranges. To help with inventory planning and warehouse management, long-term and short-term product forecasts are required. Making these forecasts adaptable is the first priority, to allow retailers to respond quickly to changes in shopper behaviour.

Traditional forecasting models couldn’t handle the scale and speed necessary that we required, so we built sophisticated deep-learning forecasts to suit the unique demands of online grocery.  

On OSP, billions of forecasts are generated every day to predict what customers will be buying. These models are trained on years of grocery data and they learn over time, understanding patterns to help you make the most of your inventory. The integration between our webshop and our supply chain results in the strongest possible forecasts, as the models leverage ecommerce inputs including in-basket sales, past sales and multiple promotional types. Our deep learning models are up to 40% more accurate than traditional retail forecasting systems.

In addition to helping with inventory management, these deep learning models feed into replenishment strategies. They automate replenishment decisions, taking this task out of the retailer’s hands and generating purchase orders automatically based on their forecasts. The AI recommends the optimal amount of stock to order to deliver on high availability and low waste.

Managing supply and demand to boost your bottom line 

A smarter forecasting system leads to a healthier sales margin, as the more stock sold, the fewer items wasted. 

OSP helps to promote the sale of products close to expiry. The intelligence of our platform understands when a product is close to expiry and launches a flash sale on the partner’s webshop to entice customers to try the product – driving profitability and reducing waste in turn.

How we’re evolving our technology for bigger and better results 

Our demand forecasting models are continuously learning and improving to help retailer partners tackle food waste with greater certainty. 

We are working on making these forecasts even more accurate, developing additional features and inputs to include in the algorithm to strengthen its predictions.